In [1]:
import matplotlib.pyplot as plt
from PIL import Image
image_path = "cover_2.png"
img = Image.open(image_path)
plt.figure(figsize=(15, 15))
plt.imshow(img)
plt.axis('off')
plt.show()
import matplotlib.pyplot as plt
from PIL import Image
image_path = "cover_2.png"
img = Image.open(image_path)
plt.figure(figsize=(15, 15))
plt.imshow(img)
plt.axis('off')
plt.show()
import pandas as pd
import seaborn as sns
import numpy as np
import sys
import geopandas as gp
import scipy.stats as stats
import plotly.express as px
sys.version
'3.8.17 (default, Jul 5 2023, 20:44:21) [MSC v.1916 64 bit (AMD64)]'
data = pd.read_csv("immo_data.csv")
# Check col
data.info()
<class 'pandas.core.frame.DataFrame'> RangeIndex: 268850 entries, 0 to 268849 Data columns (total 49 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 regio1 268850 non-null object 1 serviceCharge 261941 non-null float64 2 heatingType 223994 non-null object 3 telekomTvOffer 236231 non-null object 4 telekomHybridUploadSpeed 45020 non-null float64 5 newlyConst 268850 non-null bool 6 balcony 268850 non-null bool 7 picturecount 268850 non-null int64 8 pricetrend 267018 non-null float64 9 telekomUploadSpeed 235492 non-null float64 10 totalRent 228333 non-null float64 11 yearConstructed 211805 non-null float64 12 scoutId 268850 non-null int64 13 noParkSpaces 93052 non-null float64 14 firingTypes 211886 non-null object 15 hasKitchen 268850 non-null bool 16 geo_bln 268850 non-null object 17 cellar 268850 non-null bool 18 yearConstructedRange 211805 non-null float64 19 baseRent 268850 non-null float64 20 houseNumber 197832 non-null object 21 livingSpace 268850 non-null float64 22 geo_krs 268850 non-null object 23 condition 200361 non-null object 24 interiorQual 156185 non-null object 25 petsAllowed 154277 non-null object 26 street 268850 non-null object 27 streetPlain 197837 non-null object 28 lift 268850 non-null bool 29 baseRentRange 268850 non-null int64 30 typeOfFlat 232236 non-null object 31 geo_plz 268850 non-null int64 32 noRooms 268850 non-null float64 33 thermalChar 162344 non-null float64 34 floor 217541 non-null float64 35 numberOfFloors 171118 non-null float64 36 noRoomsRange 268850 non-null int64 37 garden 268850 non-null bool 38 livingSpaceRange 268850 non-null int64 39 regio2 268850 non-null object 40 regio3 268850 non-null object 41 description 249103 non-null object 42 facilities 215926 non-null object 43 heatingCosts 85518 non-null float64 44 energyEfficiencyClass 77787 non-null object 45 lastRefurbish 80711 non-null float64 46 electricityBasePrice 46846 non-null float64 47 electricityKwhPrice 46846 non-null float64 48 date 268850 non-null object dtypes: bool(6), float64(18), int64(6), object(19) memory usage: 89.7+ MB
# Check nan value
nan_count = data.isnull().sum()
print(nan_count)
regio1 0 serviceCharge 6909 heatingType 44856 telekomTvOffer 32619 telekomHybridUploadSpeed 223830 newlyConst 0 balcony 0 picturecount 0 pricetrend 1832 telekomUploadSpeed 33358 totalRent 40517 yearConstructed 57045 scoutId 0 noParkSpaces 175798 firingTypes 56964 hasKitchen 0 geo_bln 0 cellar 0 yearConstructedRange 57045 baseRent 0 houseNumber 71018 livingSpace 0 geo_krs 0 condition 68489 interiorQual 112665 petsAllowed 114573 street 0 streetPlain 71013 lift 0 baseRentRange 0 typeOfFlat 36614 geo_plz 0 noRooms 0 thermalChar 106506 floor 51309 numberOfFloors 97732 noRoomsRange 0 garden 0 livingSpaceRange 0 regio2 0 regio3 0 description 19747 facilities 52924 heatingCosts 183332 energyEfficiencyClass 191063 lastRefurbish 188139 electricityBasePrice 222004 electricityKwhPrice 222004 date 0 dtype: int64
# 1st Features drop based on the characteristic of column
data.drop(columns=['telekomHybridUploadSpeed','telekomUploadSpeed','description','livingSpaceRange','street',
'scoutId','facilities','geo_krs','telekomTvOffer','petsAllowed','pricetrend',
'noRoomsRange','picturecount','houseNumber','streetPlain','firingTypes','interiorQual',
'interiorQual'],inplace=True)
# 2nd features drop based on nan value
data.drop(columns=['electricityKwhPrice','electricityBasePrice','energyEfficiencyClass','lastRefurbish',
'heatingCosts','noParkSpaces','thermalChar',],inplace=True)
# Check nan value in columns
nan_count_v2 = data.isnull().sum()
print(nan_count_v2)
regio1 0 serviceCharge 6909 heatingType 44856 newlyConst 0 balcony 0 totalRent 40517 yearConstructed 57045 hasKitchen 0 geo_bln 0 cellar 0 yearConstructedRange 57045 baseRent 0 livingSpace 0 condition 68489 lift 0 baseRentRange 0 typeOfFlat 36614 geo_plz 0 noRooms 0 floor 51309 numberOfFloors 97732 garden 0 regio2 0 regio3 0 date 0 dtype: int64
# 3rd features drop low-important columns
data.drop(columns=['yearConstructedRange','numberOfFloors'],inplace=True)
# Apply replacement value 'Other' to columns that are absolutely necessary but have many nan values
columns_to_fillna = ['condition', 'heatingType', 'typeOfFlat']
for column in columns_to_fillna:
data[column].fillna('Other', inplace=True)
# Delete all rows where the main value total rent is nan (do not replace with other statistical values)
data.dropna(subset=['totalRent'],inplace=True)
# Fill NaN values in floor, yearConstructed, and serviceCharge columns with median values, respectively.
columns_to_fillna = ['floor', 'yearConstructed', 'serviceCharge']
for column in columns_to_fillna:
median_value = data[column].median()
data[column].fillna(median_value, inplace=True)
# Outlier handling
numeric_columns = data.select_dtypes(include=['int64', 'float64']).columns
def remove_outliers(column):
upper_range = column.mean() + 3 * column.std()
lower_range = column.mean() - 3 * column.std()
return (column <= upper_range) & (column >= lower_range)
# Update the data frame, leaving only rows without outliers
data = data[data[numeric_columns].apply(remove_outliers, axis=0).all(axis=1)]
nan_count_v3 = data.isnull().sum()
print(nan_count_v3)
regio1 0 serviceCharge 0 heatingType 0 newlyConst 0 balcony 0 totalRent 0 yearConstructed 0 hasKitchen 0 geo_bln 0 cellar 0 baseRent 0 livingSpace 0 condition 0 lift 0 baseRentRange 0 typeOfFlat 0 geo_plz 0 noRooms 0 floor 0 garden 0 regio2 0 regio3 0 date 0 dtype: int64
numeric_columns = [ 'totalRent', 'serviceCharge', 'newlyConst', 'yearConstructed','hasKitchen', 'cellar',
'garden', 'baseRent', 'livingSpace', 'baseRentRange', 'noRooms', 'floor']
numeric_df = data[numeric_columns]
correlation_matrix = numeric_df.corr()
cmap = sns.color_palette(["#FFD700", "#FF6666", "#808080"])
sns.heatmap(correlation_matrix, fmt='.2f', annot=True, cmap=cmap, linewidths=.6)
plt.title('Correlation Heatmap', fontsize=10)
plt.show()
# Drop the column and 'baseRentRange' that are clearly unrelated to the total rental
data.drop(['cellar', 'garden', 'floor','baseRentRange'],axis=1,inplace=True)
plt.figure(figsize=(10, 6))
sns.kdeplot(data['totalRent'], shade=True, color='skyblue')
plt.xlabel('Total Rent')
plt.ylabel('Density')
plt.title('Total Rent Density Plot')
plt.show()
C:\Users\piano\AppData\Local\Temp\ipykernel_25540\3981236610.py:2: FutureWarning: `shade` is now deprecated in favor of `fill`; setting `fill=True`. This will become an error in seaborn v0.14.0; please update your code. sns.kdeplot(data['totalRent'], shade=True, color='skyblue')
# The tail on the right is too long, so I think the data needs to be adjusted.data['totalRent'].describe()
data['totalRent'].describe()
count 226473.000000 mean 810.712646 std 585.551102 min 0.000000 25% 469.000000 50% 649.000000 75% 980.000000 max 64651.000000 Name: totalRent, dtype: float64
Q1 = data['totalRent'].quantile(0.25)
Q3 = data['totalRent'].quantile(0.75)
# IQR
IQR = Q3 - Q1
# boundary calculation
lower_bound = Q1 - 1.5 * IQR
upper_bound = 10000
# Delete data that are not in the zone
data = data[(data['totalRent'] >= lower_bound) & (data['totalRent'] <= upper_bound)]
plt.figure(figsize=(10, 6))
sns.kdeplot(data['totalRent'], shade=True, color='skyblue')
plt.xlabel('Total Rent')
plt.ylabel('Density')
plt.title('Total Rent Density Plot')
plt.show()
C:\Users\piano\AppData\Local\Temp\ipykernel_25540\3981236610.py:2: FutureWarning: `shade` is now deprecated in favor of `fill`; setting `fill=True`. This will become an error in seaborn v0.14.0; please update your code. sns.kdeplot(data['totalRent'], shade=True, color='skyblue')
plt.figure(figsize=(10, 6))
sns.kdeplot(data['totalRent'], shade=True, color='skyblue')
plt.xlabel('Total Rent')
plt.ylabel('Density')
plt.title('Total Rent Density Plot')
plt.xlim(0, 4000)
plt.show()
C:\Users\piano\AppData\Local\Temp\ipykernel_25540\3233157533.py:2: FutureWarning: `shade` is now deprecated in favor of `fill`; setting `fill=True`. This will become an error in seaborn v0.14.0; please update your code. sns.kdeplot(data['totalRent'], shade=True, color='skyblue')
data['totalRent'].describe()
count 226465.000000 mean 809.536799 std 535.817365 min 0.000000 25% 469.000000 50% 649.000000 75% 980.000000 max 9774.000000 Name: totalRent, dtype: float64
geo_data = gp.read_file("plz-5stellig.shp")
geo_data.info()
<class 'geopandas.geodataframe.GeoDataFrame'> RangeIndex: 8170 entries, 0 to 8169 Data columns (total 5 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 plz 8170 non-null object 1 note 8170 non-null object 2 einwohner 8170 non-null int64 3 qkm 8170 non-null float64 4 geometry 8170 non-null geometry dtypes: float64(1), geometry(1), int64(1), object(2) memory usage: 319.3+ KB
geo_data.head(10)
| plz | note | einwohner | qkm | geometry | |
|---|---|---|---|---|---|
| 0 | 64743 | Situation unklar, evtl. haben die Häuser Marb... | 3 | 0.082066 | POLYGON ((8.98124 49.60761, 8.98312 49.60748, ... |
| 1 | 81248 | 81248 München | 121 | 1.984763 | POLYGON ((11.40526 48.16042, 11.40655 48.15678... |
| 2 | 60315 | 60315 Frankfurt am Main (FOUR) | 0 | 0.017285 | POLYGON ((8.67254 50.11264, 8.67320 50.11266, ... |
| 3 | 99331 | 99331 Geratal | 4523 | 20.207080 | POLYGON ((10.84048 50.74605, 10.84008 50.74421... |
| 4 | 60312 | 60312 Frankfurt am Main (Omniturm) | 0 | 0.001829 | POLYGON ((8.67262 50.11164, 8.67311 50.11182, ... |
| 5 | 98694 | 98694 Ilmenau | 7028 | 80.375011 | MULTIPOLYGON (((10.84019 50.64627, 10.84409 50... |
| 6 | 39628 | 39628 Bismark | 3147 | 125.135930 | POLYGON ((11.58895 52.61676, 11.59058 52.61680... |
| 7 | 64760 | 64760 Oberzent | 10503 | 165.476886 | POLYGON ((8.88883 49.60938, 8.89054 49.60985, ... |
| 8 | 35647 | 35647 Waldsolms | 4855 | 44.716972 | MULTIPOLYGON (((8.43650 50.46646, 8.44045 50.4... |
| 9 | 31195 | 31195 Lamspringe | 5876 | 70.492464 | POLYGON ((9.89197 51.96897, 9.89831 51.96569, ... |
geo_data.plot(figsize=(5,5))
<Axes: >
odata = pd.read_csv("immo_data.csv")
subset_data = odata[['regio2', 'geo_plz']]
subset_data.to_csv("subset_immo_data.csv", index=False)
gray_color = '#989898' # Hexadecimal code!!
dark_gray_color = '#707070'
more_dark_gray_color = '#404040'
background_image = Image.open('germancity_1.jpg')
fig, ax = plt.subplots(figsize=(10, 6))
ax.imshow(background_image, extent=[-0.5, 14.5, 0, 20000], aspect='auto', alpha=0.15)
top_cities = data['regio2'].value_counts().head(15)
barplot = sns.barplot(x=top_cities.index, y=top_cities.values, color=dark_gray_color, ax=ax)
for p in barplot.patches:
ax.annotate(f'{int(p.get_height())}',
(p.get_x() + p.get_width() / 2., p.get_height()),
ha = 'center', va = 'center',
xytext = (0, 10),
textcoords = 'offset points',
fontsize=9, color=more_dark_gray_color)
plt.xticks(rotation=60, color=dark_gray_color)
plt.yticks(color=dark_gray_color)
plt.ylim(0, 14000)
plt.xlabel('City', size=15, color=more_dark_gray_color)
plt.ylabel('Number of Listings', size=15, color=more_dark_gray_color)
plt.title('Top 15 Cities with Highest Listings', size=22, color= more_dark_gray_color)
plt.show()
mean_rent_by_regio2 = data.groupby('regio2')['totalRent'].mean().reset_index()
top_15_cities = mean_rent_by_regio2.sort_values(by='totalRent', ascending=False).head(15)
background_image = Image.open('germancity_2.jpg')
fig, ax = plt.subplots(figsize=(10, 6))
ax.imshow(background_image, extent=[-0.5, 14.5, 0, 3000], aspect='auto', alpha=0.15)
barplot = sns.barplot(x='regio2', y='totalRent', data=top_15_cities, color=dark_gray_color, ax=ax)
for p in barplot.patches:
ax.annotate(f'{p.get_height():.0f}',
(p.get_x() + p.get_width() / 2., p.get_height()),
ha = 'center', va = 'center',
xytext = (0, 10),
textcoords = 'offset points',
fontsize=9, color=more_dark_gray_color)
plt.xticks(rotation=60, color=dark_gray_color)
plt.yticks(color=dark_gray_color)
plt.xlabel('City', size=15, color=more_dark_gray_color)
plt.ylabel('Average Total Rent Price', size=15, color=more_dark_gray_color)
plt.title('Top 15 Cities with Highest Average Total Rent Price / €', size=18, color= more_dark_gray_color)
plt.ylim(1000, 2000)
plt.show()
# Shapefile load
shapefile_path = 'plz-5stellig.shp'
mean_rent_by_regio2 = data.groupby('regio2')['totalRent'].mean().reset_index()
top_15_cities = mean_rent_by_regio2.sort_values(by='totalRent', ascending=False).head(24)
germany_shapes = gp.read_file(shapefile_path)
# PLZ Matching
top_cities_plz = data[data['regio2'].isin(top_15_cities['regio2'])].drop_duplicates(subset=['regio2'])
germany_shapes['plz'] = germany_shapes['plz'].astype(str)
top_cities_plz['geo_plz'] = top_cities_plz['geo_plz'].astype(str)
# Shapefile Matching
mapped_data = germany_shapes.merge(top_cities_plz, left_on='plz', right_on='geo_plz')
# Extract lat and lon from the center of goe
mapped_data['lon'] = mapped_data.geometry.apply(lambda x: x.centroid.x)
mapped_data['lat'] = mapped_data.geometry.apply(lambda x: x.centroid.y)
mapped_data = mapped_data.set_index('regio2').reindex(top_15_cities['regio2']).reset_index()
# Plotly map creation
fig = px.scatter_mapbox(
mapped_data,
lat='lat',
lon='lon',
hover_name='regio2',
zoom=5,
center={"lat": 51.1657, "lon": 10.4515},
mapbox_style="open-street-map",
size_max=20,
color='regio2',
size='totalRent'
)
fig.update_layout(
title_text="Top 24 Cities in Germany by Average Total Rent",
title_font=dict(size=20),
margin={"r": 0, "t": 40, "l": 0, "b": 0},
paper_bgcolor="rgba(245,245,245,1)")
#Legend
fig.update_layout(
legend=dict(
title=("<- Circle size = Costs"),
x=1,
y=1,
bordercolor="Black",
borderwidth=0.5
))
fig.show()
top_15_cities = (
data.groupby('regio2')['totalRent']
.mean()
.reset_index()
.sort_values(by='totalRent', ascending=False)
.head(15)
)
top_15_regio2 = top_15_cities['regio2']
top_regio2_data = data[data['regio2'].isin(top_15_regio2)]
plt.figure(figsize=(12, 6))
sns.boxplot(x='regio2', y='totalRent', data=top_regio2_data, order=top_15_regio2, palette="Set3")
plt.xticks(rotation=90)
plt.xlabel('City')
plt.ylabel('Total Rent Price')
plt.title('Total Rent Price Distribution by City Top 15')
plt.tight_layout()
plt.show()
mean_livingspace_by_regio2 = data.groupby('regio2')['livingSpace'].mean().reset_index()
top_15_cities_livingspace = (
mean_livingspace_by_regio2
.sort_values(by='livingSpace', ascending=False)
.head(15)
)
background_image = Image.open('germancity_7_apt.jpg')
fig, ax = plt.subplots(figsize=(10, 6))
ax.imshow(background_image, extent=[90, 105, -0.5, 15], aspect='auto', alpha=0.15, origin='lower')
barplot = sns.barplot(x='livingSpace', y='regio2', data=top_15_cities_livingspace, color=dark_gray_color, ax=ax)
for p in barplot.patches:
ax.annotate(f'{p.get_width():.1f}',
(p.get_width(), p.get_y() + p.get_height() / 2.),
ha = 'left', va = 'center',
xytext = (5, 0),
textcoords = 'offset points',
fontsize=9, color=more_dark_gray_color)
plt.xlabel('Average Living Space', color=more_dark_gray_color)
plt.ylabel('City', color=more_dark_gray_color)
plt.xticks(color=dark_gray_color)
plt.yticks(color=dark_gray_color)
plt.title('Top 15 Cities with Highest Average Living Space / sqm', size=18, color=more_dark_gray_color)
plt.xlim(90, 105)
plt.show()
mean_totalrent_by_heatingtype = data.groupby('heatingType')['totalRent'].mean().reset_index()
sorted_heatingtype_totalrent = mean_totalrent_by_heatingtype.sort_values(by='totalRent', ascending=False)
background_image = Image.open('germancity_10_heating.jpg')
fig, ax = plt.subplots(figsize=(10, 6))
ax.imshow(background_image, extent=[0, 2000, -15, 20], aspect='auto', alpha=0.15, origin='lower')
barplot = sns.barplot(x='totalRent', y='heatingType', data=sorted_heatingtype_totalrent, color=dark_gray_color, ax=ax)
for p in barplot.patches:
ax.annotate(f'{p.get_width():.0f}',
(p.get_width(), p.get_y() + p.get_height() / 2.),
ha = 'left', va = 'center',
xytext = (5, 0),
textcoords = 'offset points',
fontsize=9, color=more_dark_gray_color)
plt.xlabel('Average Total Rent', color=more_dark_gray_color)
plt.ylabel('Heating Type', color=more_dark_gray_color)
plt.xticks(color=dark_gray_color)
plt.yticks(color=dark_gray_color)
plt.title('Average Total Rent by Heating Type / €', size=18, color=more_dark_gray_color)
plt.xlim(400, 1500)
plt.show()
import matplotlib.pyplot as plt
import seaborn as sns
from PIL import Image
import pandas as pd
mean_totalrent_by_heatingtype = data.groupby('condition')['totalRent'].mean().reset_index()
sorted_heatingtype_totalrent = mean_totalrent_by_heatingtype.sort_values(by='totalRent', ascending=False)
background_image = Image.open('germancity_11_apt.jpg')
fig, ax = plt.subplots(figsize=(10, 6))
ax.imshow(background_image, extent=[0, 2000, -5, 12], aspect='auto', alpha=0.15, origin='lower')
barplot = sns.barplot(x='totalRent', y='condition', data=sorted_heatingtype_totalrent, color=dark_gray_color, ax=ax)
for p in barplot.patches:
ax.annotate(f'{p.get_width():.0f}',
(p.get_width(), p.get_y() + p.get_height() / 2.),
ha = 'left', va = 'center',
xytext = (5, 0),
textcoords = 'offset points',
fontsize=9, color=more_dark_gray_color)
plt.xlabel('Average Total Rent', color=more_dark_gray_color)
plt.ylabel('Condition', color=more_dark_gray_color)
plt.xticks(color= dark_gray_color)
plt.yticks(color= dark_gray_color)
plt.title('Average Total Rent by Condition / €', size=18, color=more_dark_gray_color)
plt.xlim(400, 1500)
plt.show()
# Please use %matplotlib notebook
%matplotlib notebook
from matplotlib.animation import FuncAnimation
total_rent_mean = np.mean(data['totalRent'])
base_rent_mean = np.mean(data['baseRent'])
difference = total_rent_mean - base_rent_mean
fig, ax = plt.subplots()
bar_width = 0.35
opacity = 0.8
# 1000ms / 50ms = 20 frames
pause_frames = int(5000 / 50)
light_black = (0.5, 0.5, 0.5)
light_red = (1, 0.5, 0.5)
light_gold = (1, 0.85, 0.5)
positions = np.arange(3)
bar_width = 0.95
def animate_with_percentage_and_pause(i):
if i <= 25: # First bar
ax.clear()
ax.bar(positions[0], total_rent_mean * i / 25, bar_width, alpha=opacity, color=light_black, label='Total Rent')
elif i <= 50: # Second bar
ax.clear()
ax.bar(positions[0], total_rent_mean, bar_width, alpha=opacity, color=light_black, label='Total Rent')
ax.bar(positions[1], base_rent_mean * (i - 25) / 25, bar_width, alpha=opacity, color=light_red, label='Base Rent')
elif i <= 75: # Third Bar
ax.clear()
ax.bar(positions[0], total_rent_mean, bar_width, alpha=opacity, color=light_black, label='Total Rent')
ax.bar(positions[1], base_rent_mean, bar_width, alpha=opacity, color=light_red, label='Base Rent')
ax.bar(positions[2], difference * (i - 50) / 25, bar_width, alpha=opacity, color=light_gold, label='Difference')
if i == 75: # Figure
percentage = (difference / total_rent_mean) * 100 # percentage caculation
ax.text(positions[2], difference, f'{percentage:.2f}%', ha='center', va='bottom')
ax.set_title('Additional Charge Percentage / €')
ax.set_xticks(positions)
ax.set_xticklabels(['Total Rent', 'Base Rent', 'Difference'])
# Run animation (pause for 5 seconds after displaying numbers)
ani = FuncAnimation(fig, animate_with_percentage_and_pause, frames=np.arange(0, 75 + pause_frames + 1), interval=50, repeat=True, blit=True)
plt.show()
kitchen_count = data['hasKitchen'].value_counts(normalize=True)
true_count = kitchen_count.get(True, 0)
false_count = kitchen_count.get(False, 0)
df = pd.DataFrame({'True': [true_count], 'False': [false_count]})
background_image = Image.open('germancity_12_kitchen.jpeg')
fig, ax = plt.subplots(figsize=(7, 10))
ax.imshow(background_image, extent=[0.01, 1, -0.7, 1.2], aspect='auto', alpha=0.3)
light_lilac = (200/255, 162/255, 200/255)
light_turquoise = (160/255, 220/255, 220/255)
# horizontal Stacked Bar Chart
barplot = df.plot(kind='barh', stacked=True, color=[light_lilac, light_turquoise], ax=ax, width=0.1)
# Insert text in the bars
for bar in barplot.patches:
width = bar.get_width()
label_x_pos = bar.get_x() + width / 2
text = 'Yes' if bar.get_x() < 0.1 else 'No'
percentage = f'{width:.0%}'
ax.text(label_x_pos, bar.get_y() + bar.get_height()/2, f'{text} ({percentage})',
ha='center', va='center', fontsize=12)
plt.title('Does the house have a kitchen?')
plt.xlabel('Proportion')
plt.ylabel('Has Kitchen ?!')
plt.yticks(rotation=0) # rotate y axis lable
plt.legend(title='Kitchen')
plt.xlim(0, 1.0)
plt.show()
image_path = "germany_kitchen_1.png"
img = Image.open(image_path)
new_width = 150
new_height = 150
img = img.resize((new_width, new_height), Image.LANCZOS)
plt.imshow(img)
plt.axis('off')
plt.show()
